Data-driven modelling and prediction of COVID-19 infection in India and correlation analysis of the virus transmission with socio-economic factors

Abstract

Aims: The current study attempts to model the COVID-19 outbreak in India, USA, China, Japan, Italy, Iran, Canada and Germany. The interactions of coronavirus transmission with socio-economic factors in India using the multivariate approach were also investigated. Methods: Actual cumulative infected population data from 15 February to May 15, 2020 was used for determination of parameters of a nested exponential statistical model, which were further employed for the prediction of infection. Correlation and Principal component analysis provided the relationships of coronavirus spread with socio-economic factors of different states of India using the Rstudio software. Results: Cumulative infection and spreadability rate predicted by the model was in good agreement with the actual observed data for all countries (R2 ¼ 0.985121 to 0.999635, and MD ¼ 1.2e7.76%) except Iran (R2 ¼ 0.996316, and MD ¼ 18.38%). Currently, the infection rate in India follows an upward trajectory, while other countries show a downward trend. The model claims that India is likely to witness an increased spreading rate of COVID-19 in June and July. Moreover, the flattening of the cumulative infected population is expected to be obtained in October infecting more than 12 lakhs people. Indian states with higher population were more susceptible to virus infection. Conclusions: A long-term prediction of cumulative cases, spreadability rate, pandemic peak of COVID-19 was made for India. Prediction provided by the model considering most recent data is useful for making appropriate interventions to deal with the rapidly emerging pandemic